letters to nature
..............................................................
An interplanetary shock traced
by planetary auroral storms
from the Sun to Saturn
Renée Prangé1, Laurent Pallier1, Kenneth C. Hansen2, Russ Howard3,
Angelos Vourlidas3, Régis Courtin1 & Chris Parkinson4
1
LESIA, Observatoire de Paris, 5 place Jules Janssen, 92195 Meudon, France
Department of Atmospheric, Oceanic and Space Sciences, University of Michigan,
Ann Arbor, Michigan 48109, USA
3
Naval Research Laboratory, 4555 Overlook Avenue, SW Washington, DC 20375,
USA
4
Californian Institute of Technology, Jet Propulsion Laboratory and the NASA
Astrobiology Institute, MS150-21 E. California Boulevard, Pasadena, California
91125, USA
2
of Jupiter in late 2000, where three occurrences of auroral radio
storms coincided with the arrival of interplanetary shocks at the
planet21. In the absence of any image, the nature and location of the
auroral events could not be compared with that of geomagnetic
storms.
On 7 and 8 December 2000, FUV high-spatial-resolution images
of Saturn were taken with the Hubble space telescope, HST (Fig. 1)
with an excellent view of the southern polar region. The auroral oval
morphology was very different from one day to the next. On 8
December, a narrow oval surrounding the pole at high latitude, and
assigned to large-scale currrents at the boundary between open and
closed field lines17, resembles the few Saturn aurorae observed
earlier22; this configuration may be considered as a steady-state
reference. By contrast, the aurora observed on the previous day is of
.............................................................................................................................................................................
A relationship between solar activity and aurorae on Earth was
postulated1,2 long before space probes directly detected plasma
propagating outwards from the Sun3. Violent solar eruption
events trigger interplanetary shocks4 that compress Earth’s magnetosphere, leading to increased energetic particle precipitation
into the ionosphere and subsequent auroral storms5,6. Monitoring shocks is now part of the ‘Space Weather’ forecast programme
aimed at predicting solar-activity-related environmental
hazards. The outer planets also experience aurorae, and here
we report the discovery of a strong transient polar emission on
Saturn, tentatively attributed to the passage of an interplanetary
shock—and ultimately to a series of solar coronal mass ejection
(CME) events. We could trace the shock-triggered events from
Earth, where auroral storms were recorded, to Jupiter, where the
auroral activity was strongly enhanced, and to Saturn, where it
activated the unusual polar source. This establishes that shocks
retain their properties and their ability to trigger planetary
auroral activity thoughout the Solar System. Our results also
reveal differences in the planetary auroral responses on the
passing shock, especially in their latitudinal and local time
dependences.
As for that of the Earth, outer-planet magnetospheres—the
domain where the magnetic pressure of planetary origin exceeds
the solar-wind kinetic (ram) pressure—are shaped by the solar-wind
flow and they also display permanent aurorae near the magnetic
poles at infrared, far-ultraviolet (FUV) and radio wavelengths7–11.
Similarly, when a high-ram-pressure interplanetary shock wave
reaches the planet, the magnetosphere is suddenly compressed.
Among other consequences, the strength and distribution of currents and plasma are altered, and enhanced fluxes of energetic
charged particles are expected to precipitate into the auroral ovals,
which must be strongly activated. Indeed, strong statistical correlations were found between auroral radio emissions and the solarwind density and ram pressures12–16, particularly for the Earth and
Saturn, whose magnetospheres are dominated by interactions with
the solar wind at the polar-cap boundary17, between magnetic field
lines open to the interplanetary magnetic field (IMF) and closed
field lines. The correlation is looser at Jupiter, owing to the complex
nature of its magnetosphere, partly controlled by processes of
planetary origin18–20. For the Earth, an extended set of observations
of the Sun, coordinated with measurements of the solar-wind
physical parameters (density, velocity, composition, magnetic field
strength and direction, and so on) and with various indicators of the
auroral activity (among which are radio, visible and ultraviolet
emissions), allows predictions of the ionosphere and atmosphere
response to any interplanetary shock and or change in the direction
of the IMF. This is not the case for the outer planets where there are
no such coordinated observations, except during the Cassini fly-by
78
Figure 1 Discovery of an auroral storm on Saturn. The images, in a false-orange
brightness colour code, were taken with the Space Telescope Imaging Spectrograph
(STIS) onboard HST at 11:30 UT on 7 December 2000 (a), and 10:00 UT on 8 December
(b). The exposure time is 480 s, during which Saturn rotated by ,58. A spectral range
including H2 Werner–Lyman auroral bands between ,130 and ,165 nm is isolated by
selecting the filter SRF2 in combination with the FUV MAMA solar blind detector.
Nevertheless, some solar light is reflected by the planet disc in the FUV (up to ,190 nm).
It is used to define the exact location of the planet on the detector and hence to correct
pointing inaccuracies. After fitting an oblate spheroid plus ring system to the image, a grid
of coordinates, with steps of 108 in latitude and 208 in longitude, is overlaid to determine
the coordinates of the features of interest. The accuracy of the limb-fitting procedure is
,^2 pixels in each direction. The pixel size is 0.024 £ 0.024 arcsec2 and the field of
view is 24.7 £ 24.7 arcsec2. The rotational phase of Saturn at the time of the
observations is characterized by the longitude of the sub-Earth point, or local noon
(SEP ¼ 2438 and 2838 respectively in the images used here). The north polar region is
hidden behind the ring system owing to the large tilt of Saturn’s rotation axis at that time.
Conversely, the whole region limited by the southern auroral oval is visible from Earth.
Note the similarity in geometry and the differences in nightside brightness between the
auroral oval on both days, but especially the very bright feature visible poleward of the
dawnside oval on 7 December. This feature has totally disappeared on 8 December.
©2004 Nature Publishing Group
NATURE | VOL 432 | 4 NOVEMBER 2004 | www.nature.com/nature
letters to nature
a kind never observed before: a similar oval is also present, but a very
bright feature has transiently developed inside, on polar-cap
open field lines. This suggests that we have witnessed for the
first time an auroral storm at Saturn, which may be related to an
as-yet-unknown interaction of the solar wind with Saturn’s
magnetosphere.
At that time, the Sun, the Earth, Jupiter and Saturn were nearly
aligned (Fig. 2). Hence, any interplanetary shock would successively
encounter the three planets along its radial propagation outward,
even if its angular extent were small. A three-month observing
campaign of Jupiter was also in progress, involving the Galileo
spacecraft in the jovian magnetosphere, Cassini in the nearby solar
wind, and the HST remotely imaging. On the other hand, observations of the Sun and of Earth’s aurorae are performed on a
continuous basis (for example, by the POLAR and IMAGE Earth
orbiters), together with solar-wind measurements in the near
vicinity of the Earth (for example, by the ACE and WIND spacecraft), to predict environmental hazards due to energetic particle
precipitation into the auroral ionosphere (the ‘Space Weather’
programmes). A crude kinetic retracing of the solar-wind plasma
suggests that if any shock had triggered the auroral storm on Saturn,
it should have formed in early November near the Sun. Indeed,
the LASCO coronographic spectrograph onboard the Solar and
Heliographic Observatory (SOHO) detected a series of CMEs, five
of them being directed toward the Earth over the 1–10 November
period.
Thus all the conditions except one have been met that are
required to investigate the interaction of an interplanetary shock
with the Earth, Jupiter and Saturn consecutively. Only solar-wind
measurements near Saturn were still needed to characterize the
passage of a shock directly.
We could possibly have extrapolated them from the Cassini
measurements near Jupiter, that is, 4 AU upstream (1 astronomical
unit, AU ¼ 149.5 £ 106 km) using a mere ballistic projection,
t 1 ¼ t 0 þ Dr/v sw, and correcting for the difference in heliocentric
longitude. However, this procedure, based on a purely ‘kinematic’
approach and ignoring plasma physics considerations, becomes
more and more inaccurate as distance increases.
Therefore, we developed a more self-consistent method for
obtaining the solar-wind conditions at the outer planets. We
modelled the radial evolution of the plasma with a one-dimensional
Figure 2 Geometrical configuration of the planets in late 2000. The relative heliocentric
longitudes of the Earth, Jupiter and Saturn at the time when the interplanetary shock
discussed in this study (in grey, schematic) passed each of them, are plotted. At the
beginning (the end) of the passage, the angle from the Sun–Jupiter direction to the
Sun–Earth direction is 238 (17.58) as counted clockwise. For Saturn, these angles are 158
and 98, respectively. Consequently, the code used to propagate the solar wind is
anticipated to be reasonably accurate at Jupiter, and more accurate at Saturn (see text).
NATURE | VOL 432 | 4 NOVEMBER 2004 | www.nature.com/nature
spherical magnetohydrodynamic code, using measurements made
at Earth orbit as input. It was derived from the Versatile Advection
Code23, which was modified to accept the ACE data as input at 1 AU,
the inner boundary of the simulations. Although the assumption of
spherical symmetry is not strictly valid, the comparison of the
propagated solar wind with Cassini measurements shows a good
match, particularly when the planet is close to alignment with the
Sun–Earth direction24.
The 1–10 November series of CMEs triggered a series of five
shocks, detected near the Earth (1 AU) by WIND and ACE about two
days later. During this period, the POLAR Earth orbiter recorded
correlated phases of geomagnetic storms and relaxation phases.
The code predicts that the individual shocks observed at 1 AU later
merged into a single long-duration disturbance (fast shocks can
overtake slower ones). This disturbance is predicted to have passed
the jovian magnetosphere in the 18–24 November timeframe,
coinciding with the first of the jovian radio auroral storms mentioned above21. Compared with the Cassini in-situ data, the model
shock arrives ,0.7 days too early. This small disagreement stems
almost entirely from a combination of the IMF Parker spiral
structure with the ,208 Earth–Jupiter difference in heliospheric
longitude24. For Saturn, with a longitude difference of only ,108,
this effect is expected to be smaller, and hence the predictions more
accurate.
We used the code, validated at Jupiter, to propagate the shock
further out, and we estimate that it passed Saturn between early 2
December and the morning of 8 December. Consequently, our first
HST image exhibiting the polar storm was taken near the end of the
shock interaction with Saturn’s magnetosphere, while the second
more ‘typical’ image is representative of a relaxation phase toward
normal conditions, under a quiet ambient solar wind.
Thus we have assembled the first such synoptic view of the
propagation of a CME-driven interplanetary shock from the Sun
to Saturn, from in-situ measurements at the Earth and Jupiter (1
and 5 AU), and model-propagated plasma parameters at Jupiter and
Saturn (5 and 9 AU). At each planet, a strong auroral response is
recorded, based on POLAR images at the Earth, on Cassini radio
observations at Jupiter, and on HST FUV images at Saturn. Figure 3
summarizes the observations. The solar-wind magnetic field
strength and plasma ram pressure have been plotted in consecutive
panels at the three planets. In parallel we display the last disturbed
auroral image and the first ‘relaxed’ one for the Earth and Saturn.
For Jupiter, we have instead plotted the radio output response.
We now compare the detailed characteristics of the auroral
storms at the three planets, to progress towards an understanding
of the physical processes at work during interactions between shocks
and planetary magnetospheres.
At Earth, the quiet aurora, under solar-wind control, is located on
the night side on closed magnetotail field lines. Compression of the
magnetosphere by a shock leads to dramatic shifts towards low
latitudes and brightenings of the auroral ovals (by up to three orders
of magnitudes in extreme cases), which expand in local time both
sides of midnight25 (see Fig. 3). Increased emission from the
footprint of the dayside polar cusps, where ‘direct’ entry of solarwind plasma takes place, can also be observed as a result of
reconnection events with the IMF and plasma injections26.
The Saturn ‘normal’ aurora also suggests some kind of solarwind-related local-time organization of the steady-state aurora. But
in contrast with the Earth, the asymmetry in latitude and brightness
develops between dawn and dusk17. During the storm, the auroral
oval remains almost identical in size and location, but it is globally
brighter by 50% (from 1.5 £ 1010 to 2.2 £ 1010 watts of total flux
radiated in the H and H2 FUV spectrum). The brightness distribution along the oval also varies, its midnight-to-dawn sector
(01:00–05:30 local time) becoming very active, up to 70 kR
(1 R ¼ 106 photons radiated in 4p steradians), almost four times
its next-day brightness.
©2004 Nature Publishing Group
79
letters to nature
Figure 3 Synoptic view of the CME-driven shock propagation and auroral consequences.
a, the Sun; b, the Earth; c, Jupiter; d, Saturn. Right column, solar-wind plasma
parameters as a function of distance from the Sun (0 AU ) to Saturn (9 AU ). For the CME, the
mass m and the kinetic energy E c ¼ 1/2mv 2 of the ejection event are plotted. For
the interplanetary shocks, the magnetic field strength B, and the ram pressure p ¼ rv 2.
The dashed red lines and arrows indicate the duration of the event and the date of the
images on the left. In a, thick solid lines, thick dashed lines, and thin dashed lines are for
CMEs for which the geometry allows an encounter with the planets, marginally allows it,
and does not allow it, respectively. In c, the magnetohydrodynamic-model propagated
80
shock (solid line) is compared with the local Cassini measurements (thick grey line): the
dotted vertical lines around 18 November highlight the small difference (,0.7 days)
between the shock onset from the model and from the data. Left column, SOHO images of
the solar CME (green brightness scale for EIT images of the solar disc, and blue code for
LASCO coronographic images of the neighbouring solar wind) and of the planetary auroral
responses to the shock (orange brightness scale). The Earth aurora has been recorded by
the visible/FUV imager onboard POLAR (http://eiger.physics.uiowa.edu/,vis/images/).
Because simultaneous auroral images are not available at Jupiter, we have displayed
instead the global response at radio wavelengths19.
©2004 Nature Publishing Group
NATURE | VOL 432 | 4 NOVEMBER 2004 | www.nature.com/nature
letters to nature
In addition, there is a very bright, physically disconnected, feature
inside the dawnside oval, at the footprint of polar-cap field lines. It
extends polewards up to ,788 S (,68 poleward of the oval), and is
confined to the 5:30–11:00 local time sector. Its peak brightness
(95 kR) largely exceeds that of the oval and its total auroral output,
2.4 £ 1010 watts, is comparable to the oval output despite its very
limited spatial extent.
This suggests that, in response to the passage of the shock, the
auroral oval mainly brightens at both planets on the night side, at
the footprint of magnetotail magnetic field lines. However, Saturn
does not exhibit the expansion towards lower latitudes that is typical
of the geomagnetic storm oval. As for the bright polar-cap feature, it
does not seem to have any terrestrial counterpart. It is reminiscent
of the terrestrial polar cusp, frequently observed during dayside
reconnection of closed field lines with the IMF, on or slightly
poleward of the oval, and which is detected as a variable feature
on Jupiter as well26. But, in contrast to the saturnian feature, the
terrestrial polar cusp is close to noon27.
Recent studies, however, indicate that the IMF By east–west
component (that is, dawn–dusk on Saturn) has a strong influence
on the local-time position of the cusp and may shift it towards
morning or afternoon by as much as several hours28,29. Even though
polar cusps on Earth have not been observed as far as 6:00 local time,
this is an issue that should be investigated. Unfortunately, the
direction of the IMF cannot be extrapolated as the other solarwind properties can be. Another interpretation could be that strong
reconnection at Saturn does not occur on the day side, but rather on
the morning flank of the magnetopause, where it has been suggested
that a Kelvin–Helmholtz instability triggers the saturnian auroral
radio emissions30. It will be possible to check these hypotheses with
local plasma measurements taken at Saturn by the Cassini
mission.
A
22. Trauger, J. et al. Saturn’s hydrogen aurora: Wide field and planetary camera 2 imaging from the Hubble
Space Telescope. J. Geophys. Res. 103, 20237–20244 (1998).
23. Tòth, G. Versatile Advection Code; code available at khttp://www.phys.uu.nl/,tothl (1994).
24. Hanlon, P. G. et al. On the evolution of the solar wind between 1 and 5 AU at the time of the Cassini
Jupiter flyby: Multispacecraft observations of interplanetary coronal mass ejections including the
formation of a merged interaction region. J. Geophys. Res. 109, doi:10.1029/2003JA010112 (2004).
25. Frank, L. A. & Craven, J. D. Imaging results from Dynamics Explorer. Rev. Geophys. 26, 249–283
(1988).
26. Pallier, L. & Prangé, R. Detection of the southern counterpart of the jovian northern polar cusp.
Shared properties. Geophys. Res. Lett. 31, doi:10.1029/2003GL018041 (2004).
27. Lockwood, M. et al. The ionospheric signature of flux transfer events and solar wind dynamic pressure
changes. J. Geophys. Res. 95, 17113–17135 (1990).
28. Sandholt, P. E., Farrugia, C. J., Moen, J., Cowley, S. W. H. & Lybekk, B. Dynamics of the aurora and
associated currents during a cusp bifurcated event. Geophys. Res. Lett. 25, 4313–4316 (1981).
29. Milan, S. E., Lester, M., Cowley, S. W. H. & Brittnacher, M. Dayside convection and auroral
morphology during an interval of northward interplanetary magnetic field. Ann. Geophys. 18,
436–444 (2004).
30. Galopeau, P. H. M., Zarka, P. & Le Quéau, D. Source location of Saturn’s kilometric radiation: The
Kelvin-Helmholtz instability hypothesis. J. Geophys. Res. 100, 26337–26410 (1998).
Acknowledgements This work is based partly on observations with the NASA/ESA HSTobtained
at the STScI, which is operated by the AURA, Inc. for NASA.
Competing interests statement The authors declare that they have no competing financial
interests.
Correspondence and requests for materials should be addressed to R.P. ([email protected]).
..............................................................
Optically programmable electron
spin memory using semiconductor
quantum dots
Received 14 May; accepted 27 August 2004; doi:10.1038/nature02986.
1. de Mairan, J. J. Traité Physique et Historique de l’Aurore Boréale 245–279 (Mémoire de l’Académie
Royale des Sciences, Imprimerie Royale, Paris, 1733).
2. Chapman, S. & Ferraro, V. C. A. A new theory of magnetic storms. Nature 126, 129 (1930); Terr. Magn.
Atmos. Electr. 36, 171–186 (1931); Terr. Magn. Atmos. Electr. 37, 147–156 (1932).
3. Licht, A. L. Properties of the solar wind during sunspot minimum. J. Geophys. Res. 65, 1397–1400
(1960).
4. Greenstadt, E. W. Interplanetary magnetic effects of solar flares: Explorer 18 and Pioneer 5. J. Geophys.
Res. 70, 5451–5452 (1965).
5. Snyder, C. W., Neugebauer, M. & Rao, U. R. The solar wind velocity and its correlation with cosmicray variations and with solar and geomagnetic activity. J. Geophys. Res. 68, 6361–6370 (1963).
6. Axford, W. I. The interaction between the solar wind and the earth’s magnetosphere. J. Geophys. Res.
67, 3791–3796 (1962).
7. Carr, T. D. Jupiter’s decametric rotation period and the source A emission beam. Phys. Earth Planet.
Inter. 6, 21–28 (1972).
8. Zarka, P. Auroral radio emissions at the outer planets: Observations and theories. J. Geophys. Res. 103,
20159–20194 (1998).
9. Broadfoot, A. L. et al. Extreme ultraviolet observations from Voyager 1 encounter with Jupiter. Science
204, 979–982 (1979).
10. Clarke, J. T. et al. Spatial imaging of hydrogen Lyman-alpha emission from Jupiter. Astrophys. J. 240,
696–701 (1980).
11. Broadfoot, A. L. et al. Extreme ultraviolet observations from Voyager 1 encounter with Saturn. Science
212, 206–211 (1981).
12. Gallagher, D. L. & D’Angelo, N. Correlation between solar wind parameters and auroral kilometric
radiation intensity. Geophys. Res. Lett. 8, 1087–1089 (1981).
13. Zarka, P. & Genova, F. Low-frequency jovian emission and solar wind magnetic sector structure.
Nature 306, 767–768 (1983).
14. Desch, M. D. & Barrow, C. H. Direct evidence for solar wind control of Jupiter’s hectometer
wavelength radio emission. J. Geophys. Res. 89, 6819–6823 (1984).
15. Rabl, G. K. F. Periodicities of jovian broad-band kilometric radiation observed by Ulysses. Planet.
Space Sci. 41, 1079–1083 (1993).
16. Desch, M. D. & Rucker, H. O. The relationship between Saturn kilometric radiation and the solar
wind. J. Geophys. Res. 88, 8999–9006 (1983).
17. Cowley, S. H. W., Bunce, E. J. & Prangé, R. Saturn’s polar ionospheric flows and their relation to the
main auroral oval. Ann. Geophys. 22, 2685–2688 (2004).
18. Vasyliunas, V. M. in Physics of the Jovian Magnetosphere (ed. Dessler, A. J.) 395–453 (Cambridge Univ.
Press, Cambridge, UK, 1990).
19. Prangé, R. et al. Detailed study of FUV Jovian auroral features with the post COSTAR Hubble Faint
Object Camera. J. Geophys. Res. 103, 20195–20210 (1998).
20. Pallier, L. & Prangé, R. More about the structure of the high latitude Jovian aurorae. Planet. Space Sci.
49, 1159–1173 (2001).
21. Gurnett, D. A. et al. Control of Jupiter’s radio emission and aurorae by the solar wind. Nature 415,
985–987 (2002).
NATURE | VOL 432 | 4 NOVEMBER 2004 | www.nature.com/nature
Miro Kroutvar, Yann Ducommun, Dominik Heiss, Max Bichler,
Dieter Schuh, Gerhard Abstreiter & Jonathan J. Finley
Walter Schottky Institut, Technische Universität München, Am Coulombwall 3,
D-85748 Garching, Germany
.............................................................................................................................................................................
The spin of a single electron subject to a static magnetic field
provides a natural two-level system that is suitable for use as a
quantum bit, the fundamental logical unit in a quantum computer1–3. Semiconductor quantum dots fabricated by strain
driven self-assembly4 are particularly attractive for the realization
of spin quantum bits, as they can be controllably positioned5,
electronically coupled6 and embedded into active devices7–10. It
has been predicted that the atomic-like electronic structure4 of
such quantum dots suppresses coupling of the spin to the solidstate quantum dot environment11–14, thus protecting the ‘spin’
quantum information against decoherence15,16. Here we demonstrate a single electron spin memory device in which the electron
spin can be programmed by frequency selective optical excitation. We use the device to prepare single electron spins in
semiconductor quantum dots with a well defined orientation,
and directly measure the intrinsic spin flip time and its dependence on magnetic field. A very long spin lifetime is obtained,
with a lower limit of about 20 milliseconds at a magnetic field of
4 tesla and at 1 kelvin.
We begin by summarizing the operating principles of our optical
spin storage device before presenting the measurements of the spin
flip time, its dependence on magnetic field and the determination of
the underlying mechanism. The structure of the devices investigated
and the measurement techniques are summarized in Fig. 1. The
samples consist of a single layer of self-assembled Ga(In)As quantum dots (QDs) embedded within the intrinsic region of a p-type
GaAs Schottky photodiode17. Single electron–hole pair excitations
(excitons) are generated directly in QD ground states by frequency-
©2004 Nature Publishing Group
81